Yoshii et al. Cardiovasc Diabetol (2019) 18:85 https://doi.org/10.1186/s12933-019-0889-y Cardiovascular Diabetology

ORIGINAL INVESTIGATION Open Access Cardiac ischemia–reperfusion injury under insulin‑resistant conditions: SGLT1 but not SGLT2 plays a compensatory protective role in diet‑induced obesity Akira Yoshii, Tomohisa Nagoshi* , Yusuke Kashiwagi, Haruka Kimura, Yoshiro Tanaka, Yuhei Oi, Keiichi Ito, Takuya Yoshino, Toshikazu D. Tanaka and Michihiro Yoshimura

Abstract Background: Recent large-scale clinical trials have shown that SGLT2-inhibitors reduce cardiovascular events in diabetic patients. However, the regulation and functional role of cardiac sodium–glucose cotransporter (SGLT1 is the dominant isoform) compared with those of other glucose transporters (insulin-dependent GLUT4 is the major isoform) remain incompletely understood. Given that glucose is an important preferential substrate for myocardial energy metabolism under conditions of ischemia–reperfusion injury (IRI), we hypothesized that SGLT1 contributes to cardioprotection during the acute phase of IRI via enhanced glucose transport, particularly in insulin-resistant phenotypes. Methods and results: The hearts from mice fed a high-fat diet (HFD) for 12 weeks or a normal-fat diet (NFD) were perfused with either the non-selective SGLT-inhibitor phlorizin or selective SGLT2-inhibitors (tofoglifozin, ipragli- fozin, canaglifozin) during IRI using Langendorf model. After ischemia–reperfusion, HFD impaired left ventricular developed pressure (LVDP) recovery compared with the fndings in NFD. Although phlorizin-perfusion impaired LVDP recovery in NFD, a further impaired LVDP recovery and a dramatically increased infarct size were observed in HFD with phlorizin-perfusion. Meanwhile, none of the SGLT2-inhibitors signifcantly afected cardiac function or myocardial injury after ischemia–reperfusion under either diet condition. The plasma membrane expression of GLUT4 was sig- nifcantly increased after IRI in NFD but was substantially attenuated in HFD, the latter of which was associated with a signifcant reduction in myocardial glucose uptake. In contrast, SGLT1 expression at the plasma membrane remained constant during IRI, regardless of the diet condition, whereas SGLT2 was not detected in the hearts of any mice. Of note, phlorizin considerably reduced myocardial glucose uptake after IRI, particularly in HFD. Conclusions: Cardiac SGLT1 but not SGLT2 plays a compensatory protective role during the acute phase of IRI via enhanced glucose uptake, particularly under insulin-resistant conditions, in which IRI-induced GLUT4 upregulation is compromised. Keywords: SGLT1, Phlorizin, SGLT2-inhibitor, GLUT4, Insulin resistance, Ischemia–reperfusion injury, Glucose uptake

*Correspondence: [email protected] Division of Cardiology, Department of Internal Medicine, The Jikei University School of Medicine, 3‑25‑8, Nishi‑Shinbashi, Minato‑ku, Tokyo 105‑8461, Japan

© The Author(s) 2019. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Yoshii et al. Cardiovasc Diabetol (2019) 18:85 Page 2 of 14

Background evaluated based on either the mRNA or protein expres- Although fatty acids are the predominant fuel of energy sion from whole heart tissues in the relatively chronic metabolism in the normal adult heart, glucose is an phase. We also recently showed that SGLT1 was highly important preferential substrate under specifc patho- expressed in the plasma membrane fraction of the logical conditions, such as ischemia–reperfusion injury human autopsied heart as well as the murine Langendorf (IRI), as it provides greater efciency for producing high- perfused heart, although the transmembrane protein energy products per oxygen molecule consumed than expression was not signifcantly afected, at least during fatty acids [1–6]. During the acute phase of IRI, under the acute phase of IRI [7]. Subsequently, using a murine conditions of impaired mitochondrial oxidative phospho- Langendorf model, we reported that the non-selective rylation, increased glycolysis becomes a major mecha- SGLT-inhibitor phlorizin perfusion predisposes the heart nism by which the heart maintains ATP generation and to profound IRI, which was associated with a signifcant is critical for the cardiomyocyte survival [2]. Terefore, decrease in the cardiac tissue ATP content, as a result of targeting the acceleration of myocardial glucose utiliza- a reduction in the glucose uptake as well as the glycolytic tion during the acute phase of IRI is a potential therapeu- fux in the heart during ischemia–reperfusion [7]. Tese tic strategy for protecting cardiomyocyte and improving data suggest that cardiac SGLT1 signifcantly contributes the cardiac functional recovery [2, 5, 7, 8]. Tis approach to cardioprotection against IRI by replenishing the ATP may become of particular importance under insulin- stores in ischemic cardiac tissues through promoting glu- resistant conditions, such as diabetes mellitus, in which cose utilization. However, the regulation and functional the glucose utilization is further impaired in response to signifcance of cardiac SGLT1 under insulin-resistant various stimuli, including insulin stimulation as well as conditions, in which the GLUT4 expression is decreased, ischemic insult. remain poorly understood. Glucose utilization is initiated by the uptake of glu- Recent studies have shown that chronic excessive cose via glucose transporters, which appears to be the actions of SGLT1 actually have a detrimental impact rate-limiting step in glycolytic fux in the heart [5, 7, 9]. on the heart [20, 22–24]. Considering that SGLT1 was Glucose transporters are divided into two major families: reported to be chronically over-activated in the diabetic facilitated glucose transporters (GLUTs) and sodium- heart [19, 21, 25], the question remains whether cardiac coupled active transporters (SGLTs) [6, 7]. Among the SGLT1 detrimentally impacts the diabetic myocardium, 12 subtypes of GLUTs conserved in mammals, GLUT1 or exerts a protective efect against ischemic insult as a and GLUT4 appear to be the major glucose transport- compensatory mechanism for compromised GLUT4 ers in the heart, and GLUT4 is thought to be the most under insulin-resistant conditions. abundant, accounting for approximately 70% of glucose To better understand the role and functional sig- transporters [6, 9, 10]. GLUT4 resides mainly in the nifcance of cardiac SGLT1 during IRI under insu- intracellular vesicles under basal conditions and is trans- lin-resistant conditions, we studied the responses of located to the plasma membrane in response to insulin phlorizin-perfused mouse hearts to IRI using high-fat as well as other pathological processes, such as ischemic diet (HFD)-induced obese mice. insult. It was previously reported that GLUT4-mediated enhanced glucose transport represents an essential pro- Methods tective mechanism against IRI and other pathological Animal models stresses [2, 8, 9, 11, 12]. However, it was also reported All animal procedures conformed to the National Insti- that the cardiac GLUT4 expression is decreased under tutes of Health Guide for the Care and Use of Labora- insulin-resistant conditions, such as diabetes, in asso- tory Animals and were approved by the Animal Research ciation with the reduction in glucose uptake, leading to Committee at the Jikei University School of Medicine impaired glucose utilization in the heart [10, 13, 14]. (2016-038C4). Male C57BL/6 mice at 8 weeks of age Although the regulation and the functional roles of were fed either a HFD—consisting of 60% calories from GLUTs in the heart have been intensively investigated in fat, 20% from protein, and 20% from carbohydrates a variety of in vitro and in vivo models [2, 8, 9, 12–15], (D12492; Research Diets, New Brunswick, NJ, USA)—or less is known about the role and functional signifcance a matched control normal-fat diet (NFD)—consisting of of SGLTs in the heart, particularly under insulin-resistant 10% calories from fat, 20% from protein, and 70% from conditions. In the heart, SGLT1, not SGLT2, is consid- carbohydrates (D12450J; Research Diets)—for 12 weeks. ered to be the dominant isoform [16–18]. Previous stud- To investigate the dose efects of SGLT2-inhibitors on ies showed that SGLT1 is highly expressed in the heart, myocardial IRI, 10-week-old male ICR mice that had especially under conditions of ischemia [18, 19], hyper- been fed normal chow (CE-2; CLEA Japan, Tokyo, Japan) trophy [18, 20], or diabetes [19, 21], all of which were were used. Yoshii et al. Cardiovasc Diabetol (2019) 18:85 Page 3 of 14

Glucose and insulin tolerance tests ischemia–reperfusion, as previously described [7, 11]. For the glucose tolerance test, mice were fasted for 16 h, Briefy, hearts were sectioned from apex to base into four anaesthetized (isofurane, 1%, inhalation), and 1 mg/g 2-mm sections. To delineate the infarct, sections were body weight of glucose was injected intraperitoneally incubated in 5% (wt/vol) TTC in PBS at 37 °C for 20 min. (i.p.). For the insulin tolerance test, mice were fasted for For each section, the area at risk (AAR) and infarct area 2 h, and 1 U/kg body weight of insulin was injected i.p. In were measured from enlarged digital micro-graphs. Per- both tests, blood samples were collected at baseline and cent myocardial infarction (%MI) was calculated as the at 15, 30, 60, 90, and 120 min after injection, and plasma total infarction area divided by the total AAR for that glucose concentrations were measured using a commer- heart. cially available glucose meter (MEDISAFE FIT; Terumo Corporation, Tokyo, Japan). Glucose uptake in Langendorf hearts After 12 h of fasting, the rate of glucose transport in Lan- Experiments in Langendorf hearts gendorf perfused hearts was measured by detection of After 8 to 12 h of fasting, mice were heparinized the amount of 2-deoxy-d-glucose (2-DG) using a 2-DG (1000 IU/kg, i.p.) and anesthetized (pentobarbital, 60 mg/ Uptake Measurement kit (Cosmo Bio Co., Ltd., Tokyo, kg, i.p.) in order to eliminate sufering. Te heart was Japan) according to the same protocol as previously then rapidly excised, and the aorta was cannulated onto described [7]. a Langendorf apparatus, followed by retrograde-per- fusion at a constant pressure (80 mmHg) with modifed Cardiac enzyme measurement Krebs–Henseleit bufer (11 mM glucose, 118 mM NaCl, Creatine phosphokinase (CPK) levels were measured in 4.7 mM KCl, 2.0 mM CaCl­ 2, 1.2 mM MgSO­ 4, 1.2 mM the efuent using an enzymatic activity assay, as previ- KH­ 2PO4, 25 mM ­NaHCO3, 0.5 mM EDTA), as previously ously described [7, 26]. Values were corrected for coro- described [7, 11, 26]. A temperature-regulated heart nary fow and heart weight [CPK (U/l) × coronary fow chamber was then placed around the heart in order to (l/min)/heart weight (g) = U/min/g]. keep the perfused heart at a certain temperature. Cardiac hemodynamics was measured using a water-flled bal- Cardiac muscle fractionation loon catheter introduced into the left ventricle. Te preparation and fractionation of plasma membranes from cardiac muscles was performed using Minute Ischemia–reperfusion model Plasma Membrane Protein Isolation Kit (Invent Biotech- nologies, Eden Prairie, MN) according to the manufac- After a stabilization period of 20 min, the control hearts ture’s protocol as previously described [7]. After plasma were perfused for another 10 min before ischemia–reper- membrane fractionation, the protein concentrations were fusion in order to measure the baseline pre-ischemia car- determined according to a Bradford assay, and equal diac function. Subsequently, global ischemia was applied amounts of protein were loaded for immunoblotting. by eliminating fow followed by reperfusion, as previously −4 described [7]. Where indicated, ­10 mol/L of phlorizin Immunoblotting (Sigma-Aldrich, Tokyo, Japan) used as a non-specifc SGLT-inhibitor, 5 or 50 µmol/L of tofoglifozin (kindly Immunoblotting was performed as previously described provided by Sanof K.K., Tokyo, Japan), 10 µmol/L of [7, 26–28] with rabbit polyclonal anti-SGLT1 (1:200, ipraglifozin (kindly provided by Astellas Pharma Inc., #H-85; Santa Cruz Biotechnology, CA, USA), anti- Ibaraki, Japan), or 15 µmol/L of canaglifozin (kindly pro- GLUT4 (1:1000, #ab33780; Abcam, Cambridge, MA, vided by Mitsubishi Tanabe Pharma Co., Osaka, Japan), USA), anti-GLUT1 (1:1000, #12939; Cell Signaling Tech- all of which were used as specifc SGLT2-inhibitors, was nology, Tokyo, Japan), or mouse monoclonal anti-α1 Na– added to the bufer during the pre-ischemia perfusion K-ATPase (1:1000, #ab7671; Abcam). Te signals were and reperfusion periods. Phlorizin and SGLT2-inhibi- detected using chemiluminescence. tors were dissolved in dimethylsulfoxide (DMSO), and RNA isolation, reverse transcription (RT) and real‑time the solvent concentration (0.1% DMSO) was identically polymerase chain reaction (PCR) maintained in the control group. For the immunoblot- ting analysis, the individual perfused hearts were snap Total RNA was extracted from the frozen tissues using TRIzol reagent (Invitrogen) and a quantitative real-time frozen in liquid nitrogen and stored at − 80 °C prior to protein extraction. Triphenyltetrazolium chloride (TTC) PCR was performed using a StepOnePlus Real-time PCR staining was performed to determine the myocardial System and the StepOne Software program (Applied Bio- infarct size using the individual perfused hearts after systems), as described previously [28]. Te real-time PCR protocol consisted of one cycle at 95 °C for 20 s followed Yoshii et al. Cardiovasc Diabetol (2019) 18:85 Page 4 of 14

by 40 cycles at 95 °C for 1 s and 60 °C for 20 s using the were higher in HFD mice than in NFD mice (Fig. 1c). Te primers for Slc5a2 (Mm00453831_m1; Applied Biosys- glucose tolerance test (Fig. 1c) and insulin tolerance test tems) and GAPDH (Mn03302249_g1; Applied Biosys- (Fig. 1d) clearly demonstrated that 12-week HFD feeding tems). Te transcriptional levels were determined using induced glucose intolerance and insulin resistance. the ΔΔCt method with normalization to GAPDH. Efects of phlorizin on the cardiac function during IRI Statistical analysis in HFD mice Te data are presented as the mean ± standard error of Te experimental protocols of ex vivo Langendorf the mean of at least three independent experiments. Stu- ischemia–reperfusion (Fig. 2a) and the representative dent’s t-test was used for the comparison of two data sets. tracing of left ventricular developed pressure (LVDP) Te hemodynamic parameters, %MI determined accord- during IRI (Fig. 2b) are shown. Te baseline car- ing to TTC staining and 2-DG uptake were compared diac function measured at the end of the 10-min pre- using one-way analysis of variance (AONVA) followed ischemia perfusion period indicated that HFD mice had by post hoc Bonferroni or Turkey tests. Te CPK levels increased contractile activity, consistent with the previ- were compared using Kruskal–Wallis test and Mann– ous studies (Table 1) [29, 30]. Phlorizin-perfusion did Whitney U test. A value of P < 0.05 was considered to be not signifcantly afect the baseline cardiac function in signifcant. hearts of either NFD or HFD mice. After 30-min global ischemia followed by 40-min reperfusion, HFD per Results se signifcantly reduced LVDP recovery (22.5% ± 3.5% Efects of 12‑week HFD feeding versus 68.3% ± 7.1% recovery from baseline, P < 0.01, After 12 weeks of HFD feeding, mice developed marked Fig. 2c, d) and the rate pressure product (RPP), cal- obesity with a 44% increase in body weight compared culated as LVDP × heart rate, in order to consider with NFD mice (Fig. 1a, b). Fasting plasma glucose levels the impact of the heart rate on the cardiac function

Fig. 1 Twelve-week HFD feeding induced obesity, glucose intolerance, and insulin resistance in mice. a Appearance of the obese mice after 12 weeks of HFD feeding. Bars 1 cm. b Body weight changes during HFD feeding (n 6 each). c Plasma glucose levels during glucose tolerance = = tests (NFD; n 9, HFD; n 12). d Plasma glucose levels during insulin tolerance tests (n 12 each). **P < 0.01 versus NFD at each time point. NFD = = = normal-fat diet, HFD high-fat diet Yoshii et al. Cardiovasc Diabetol (2019) 18:85 Page 5 of 14

Fig. 2 Cardiac functional recovery after ischemia–reperfusion injury with or without phlorizin-perfusion. a Experimental protocols indicating the duration and time course of ischemia/reperfusion are shown. The CPK activity released into the perfusate was measured at the point in the protocol indicated by the solid arrows. b Representative tracing of LVDP during ischemia–reperfusion with or without phlorizin-perfusion. LVDP profles (c) and LVDP recovery (percent of baseline) (d) measured at the indicated time points during ischemia–reperfusion in NFD (open black square; n 6), = NFD phlorizin-perfused (flled black square; n 6), HFD (open pink square; n 8), and HFD phlorizin-perfused (flled pink square; n 6) hearts are = = = shown. *P < 0.05 and **P < 0.01 versus the NFD group at each time point; †P < 0.05 and ‡P < 0.01 versus the HFD group at each time point. SGLT2i SGLT2-inhibitors, min minutes, LVDP left ventricular developed pressure, I ischemia, R reperfusion

Table 1 Baseline cardiac function of ex vivo perfused hearts with or without phlorizin-perfusion NFD NFD-phlorizin HFD HFD-phlorizin (n 6) (n 6) (n 8) (n 8) = = = = LVSP, mmHg 98.9 5.1 112 8.5 119 4.9* 133 11* ± ± ± ± LVEDP, mmHg 9.2 0.4 10 0.8 9.6 0.2 9.4 0.5 ± ± ± ± LVDP, mmHg 89.8 5.1 102 8.9 109 5.0* 123 11* ± ± ± ± dp/dt, mmHg/s 2742 247 3277 308 3478 162 4098 463 + ± ± ± ± dp/dt, mmHg/s 2308 200 2450 277 2814 177 3380 518 − − ± − ± − ± − ± HR, bpm 272 44 306 35 266 27 284 22 ± ± ± ± RPP, mmHg bpm 25,429 5192 31,251 4897 28,650 2714 35,970 5245 ± ± ± ± Coronary fow, mL/min 3.30 0.47 2.93 0.36 3.19 0.32 3.58 0.38 ± ± ± ± NFD normal fat diet, HFD high fat diet, LVSP left ventricular systolic pressure, LVEDP left ventricular end-diastolic pressure, LVDP left ventricular developed pressure, HR heart rate, RPP rate pressure product *P < 0.05 versus NFD group

pressure (LVEDP) (Additional fle 1: Fig. S1B) and the (6733 ± 1271 versus 20,610 ± 4716 mmHg bpm at the end of IRI, P < 0.01, Additional fle 1: Fig. S1A) in asso- signifcant impairment of the maximum rate of con- ciation with an increase in left ventricular end-diastolic traction (+dp/dtmax) and maximum rate of relaxation Yoshii et al. Cardiovasc Diabetol (2019) 18:85 Page 6 of 14

the CPK area under the curve (AUC) for the CPK level, (−dp/dtmin) (Additional fle 1: Fig. S1C) compared with those values in NFD hearts. Although phlorizin- was also shown (Fig. 3b) [7, 11]. During the baseline pre- perfusion also impaired the cardiac functional recov- ischemia perfusion period, no CPK activity was detecta- ble. After IRI, the CPK activity was signifcantly increased ery in NFD hearts (LVDP recovery rate: 29.5% ± 4.5%, in HFD hearts compared with NFD hearts (CPK-AUC: P < 0.01, RPP: 9102 ± 867 mmHg bpm, P < 0.05), con- sistent with our previous study [7], further dramatically 5.2 ± 1.7 versus 1.3 ± 0.2 U/g, P < 0.05). Although phlori- impaired cardiac functional recovery was observed in zin-perfusion also increased CPK release in NFD hearts HFD hearts with phlorizin-perfusion (LVDP recovery (8.3 ± 1.7 U/g, P < 0.01), consistent with our previous study [7], further dramatically increased CPK release rate: 7.1% ± 1.9%, RPP 2146 ± 601 mmHg bpm, P < 0.01 versus HFD without phlorizin-perfusion). Tese data was observed in HFD hearts with phlorizin-perfusion suggest that SGLT1 is vital for the cardiac functional (23.2 ± 6.9 U/g, P < 0.05 versus HFD without phlorizin- recovery after ischemia–reperfusion, particularly under perfusion). Tese fndings correlated with larger infarcts insulin-resistant conditions. in NFD hearts with phlorizin-perfusion and HFD hearts (52.5% ± 3.5% [P < 0.01] and 60.2% ± 1.5% [P < 0.01], ver- Cardiac injury after ischemia–reperfusion sus 32.3% ± 2.2% [NFD], respectively), which was par- Te activity of CPK released into the perfusate during ticularly striking in HFD hearts with phlorizin-perfusion reperfusion was measured as an index of myocardial (71.8% ± 4.0%, P < 0.05 versus HFD hearts without phlo- injury (Fig. 3a), and the total amount of CPK released rizin-perfusion), compared with those noted in the NFD during the entire reperfusion period, as indicated by hearts (Fig. 3c, d).

Fig. 3 Myocardial injury after ischemia–reperfusion with or without phlorizin-perfusion. a CPK profles in the efuent collected during the reperfusion period. b The area under the curve (AUC) was calculated from the CPK profle shown in a (NFD; n 11, NFD with phlorizin-perfusion; = n 7, HFD; n 10, HFD with phlorizin-perfusion; n 4). *P < 0.05 and **P < 0.01 versus the NFD group at each time point; †P < 0.05 and ‡P < 0.01 = = = versus the HFD group at each time point. c Micrograph showing representative TTC staining of cardiac sections obtained from NFD, NFD with phlorizin-perfusion, HFD, and HFD with phlorizin-perfusion groups after ischemia–reperfusion. Bars 3 mm. d Efects on the quantitated = cumulative infarct area size in NFD (n 9), NFD phlorizin-perfused (n 6), HFD (n 5), and HFD phlorizin-perfused (n 4) hearts. **P < 0.01 versus = = = = the NFD group; †P < 0.05 versus the HFD group. %MI myocardial infarct area/ventricular area Yoshii et al. Cardiovasc Diabetol (2019) 18:85 Page 7 of 14

Direct efects of selective SGLT2‑inhibitors on either NFD diet conditions (Additional fle 1: Fig. S3). Likewise, the or HFD hearts during IRI regulation of the expression of those glucose transport- Increasing attention has been focused on SGLT2-inhib- ers was not afected by phlorizin-perfusion (Additional itors as novel anti-diabetic agents that reduce cardio- fle 1: Fig. S4), indicating that SGLT1 inhibition negligibly vascular events, as shown in recent large clinical trials infuences the regulation of GLUT4 and GLUT1 expres- [31–33]. We therefore next investigated the direct efects sion in NFD and HFD hearts, both at baseline and after of selective SGLT2-inhibitors on the cardiac functional IRI. In addition, these data may rule out the possibility of recovery and myocardial injury after ischemia–reperfu- the non-specifc inhibitory efects of phlorizin on other sion in both HFD and NFD hearts. glucose transporters, at least in the present experimental No SGLT2 mRNA expression was detected even in model. HFD mouse hearts (Additional fle 1: Fig. S2). Each SGLT2-inhibitor—namely 5 µM of tofoglifozin, 10 µM Glucose utilization in the hearts during IRI of ipraglifozin, or 15 µM of canaglifozin—was perfused To test the functional signifcance of the contrasting according to the protocol of Langendorf ischemia–rep- changes in the plasma membrane expression of glu- erfusion using HFD hearts (Fig. 2a), and no signifcant cose transporters, we measured the myocardial glucose efects of these SGLT2-inhibitors on the baseline cardiac uptake (one of the major determinants of myocardial function were noted (Additional fle 1: Table S1). After glucose utilization) before and after IRI [7]. At baseline, IRI, none of the SGLT2-inhibitors signifcantly afected there were no signifcant between-group diferences the cardiac function, including LVDP recovery after IRI, in the glucose uptake rate (Fig. 5d). In contrast, during or CPK release in HFD mice (Fig. 4a–d). ischemia–reperfusion, the cardiac glucose uptake was To simply assess the dose efects of SGLT2-inhibitors, dramatically enhanced by the ischemic insult in NFD either 5 or 50 µM of tofoglifozin was perfused during hearts, which was signifcantly attenuated in HFD hearts. 20-min global ischemia followed by 40-min reperfu- Tis was thought to be mainly due to the blunted GLUT4 sion in ICR mice fed normal chow. Te baseline cardiac reaction to IRI. Of note, phlorizin considerably reduced function was not afected by either dose of tofoglifozin the glucose uptake after IRI, particularly in HFD hearts (Additional fle 1: Table S2). However, while 5 µM of compared to the other subjects (Fig. 5d). Tese data sug- tofoglifozin did not afect the cardiac function or CPK gest that SGLT1 plays a critical role in myocardial glucose release during IRI, 50 µM of tofoglifozin signifcantly utilization during IRI, particularly under insulin-resistant reduced LVDP recovery in association with an increased conditions. CPK release after IRI (Fig. 4e–h), similar to the efects observed under phlorizin-perfusion [7]. Taken together, Discussion these fndings indicated no harmful efects of selec- In the present study using diet-induced obesity model, tive SGLT2-inhibitors during IRI, even in HFD hearts, we showed that the inhibition of cardiac SGLT1 by phlo- although detrimental efects were noted at supra-phar- rizin during the acute phase of IRI led to an impaired car- macological doses, possibly through non-selective SGLT1 diac functional recovery and increased myocardial injury, inhibition. which were associated with signifcant reductions in the myocardial glucose uptake enhanced by IRI. Te novel The expression of glucose transporters in HFD hearts fndings of the present study are as follows: (i) the detri- during IRI mental infuence of phlorizin during IRI described above In order to clarify the mechanisms underlying these det- was more striking in subjects with diet-induced obesity rimental efects of SGLT1 inhibition during IRI, particu- than in normal subjects. (ii) Tis mechanistic connection larly in HFD hearts, we next examined the dynamics of was demonstrated by the blunted response of GLUT4 glucose transporters in these murine hearts. Immunob- upregulation to IRI in contrast to the constant expres- lotting with plasma membrane fractionation revealed that sion of SGLT1 in diet-induced obesity mouse hearts. the GLUT4 expression was signifcantly reduced in HFD (iii) In contrast to phlorizin, selective SGLT2-inhibitors hearts compared with NFD hearts at baseline (Fig. 5a, did not signifcantly afect the cardiac functional recov- c). Te GLUT4 expression at the plasma membrane was ery or myocardial injury after ischemia–reperfusion, signifcantly increased after IRI in NFD hearts, while the even in diet-induced obesity mouse hearts, in which the response of GLUT4 to IRI was substantially attenuated in SGLT2 expression was not detected. Tese data indicate HFD hearts (Fig. 5b, c). In contrast, the SGLT1 expres- the increasing reliance on SGLT1 for cardiac functional sion at the plasma membrane remained constant during recovery and energy metabolism during IRI in insulin- IRI, regardless of diet conditions (Fig. 5a–c). Te GLUT1 resistant diabetic phenotypes, in which IRI-induced expression was also equivalent during IRI under both GLUT4 upregulation is compromised. Yoshii et al. Cardiovasc Diabetol (2019) 18:85 Page 8 of 14

Fig. 4 Cardiac functional recovery and myocardial injury after ischemia–reperfusion injury with or without SGLT2-inhibitor perfusion. LVDP profles and myocardial injury in HFD (a–d) and NFD (e–h) mice are shown. LVDP profles (a) and LVDP recovery (percent of baseline) (b) measured at the indicated time points during ischemia–reperfusion in the control (open pink square; n 8, as shown in Fig. 2c, d), 5 µM tofoglifozin-perfused (blue = square; n 5), 10 µM ipraglifozin-perfused (green square; n 6), and 15 µM canaglifozin-perfused (yellow square; n 8) hearts are shown. CPK = = = profles in the efuent collected during the reperfusion period (c) and the area under the curve (AUC) (d) calculated from the CPK profle (c) are shown (control: n 5, as shown in Fig. 3a, b; tofoglifozin perfusion: n 5; ipraglifozin perfusion: n 7; canaglifozin perfusion: n 7). **P < 0.01 = = = = versus the HFD group at each time point. LVDP profles (e) and LVDP recovery (percent of baseline) (f) measured at the indicated time points during ischemia–reperfusion in the control (open black square; n 10), 5 µM tofoglifozin-perfused (blue square; n 8), and 50 µM tofoglifozin-perfused = = (purple square; n 8) hearts are shown. CPK profles in the efuent collected during the reperfusion period (g) and the AUC (h) calculated from the = CPK profle (g) are shown (control: n 12; tofoglifozin [5 µM]-perfusion: n 7; tofoglifozin [50 µM]-perfusion: n 7). *P < 0.05 and **P < 0.01 versus = = = control at each time point. Tofo tofoglifozin, Ipra ipraglifozin, Cana canaglifozin

Various mechanisms underlying the detrimental structure, and energy metabolism have been intensively infuence of hyperglycemia under diabetic and/or insu- investigated [34–37]. One key point to note in the pre- lin-resistant conditions on the myocardial function, sent isolated heart perfusion study is that the detrimental Yoshii et al. Cardiovasc Diabetol (2019) 18:85 Page 9 of 14

Fig. 5 The expression of glucose transporters and the glucose uptake in myocardium during ischemia–reperfusion. Representative immunoblots of GLUT4 and SGLT1 in the plasma membrane fraction from the murine perfused hearts at baseline measured at the end of 10-min pre-ischemia perfusion (a) and before and after IRI (b) are shown. c Densitometric quantitation normalized to the level of either the GLUT4 or SGLT1 expression in NFD hearts before IRI are shown ([GLUT4] NFD or HFD without IRI: n 8 each, NFD or HFD with IRI: n 6 each; [SGLT1] NFD or HFD without IRI: = = n 9 each, NFD or HFD with IRI: n 7 or n 6, respectively). *P < 0.05, **P < 0.01 versus NFD hearts before IRI; ‡P < 0.01 versus NFD hearts after IRI. = = = In both a and b, immunoblots of ­Na+/K+ ATPase from the same membrane are shown as a loading control for the membrane fraction. d Glucose uptake in NFD (open black square), NFD phlorizin-perfused (flled black square), HFD (open pink square), and HFD phlorizin-perfused (flled pink square) hearts under the pre-ischemic baseline conditions and post-ischemic condition measured after 20-min reperfusion following 30-min global ischemia (baseline/post-ischemic condition, NFD: n 4/n 5, NFD with phlorizin-perfusion: n 3/n 5, HFD: n 7/n 7, and HFD with ======phlorizin-perfusion: n 3/n 7) *P < 0.05 and **P < 0.01 versus NFD hearts under post-ischemic conditions, †P < 0.05 versus HFD hearts under = = post-ischemic conditions, $$P < 0.01 versus corresponding controls at baseline efects of HFD on these parameters became evident recovery and increased myocardial injury after ischemia– after ischemia–reperfusion, not being noted at baseline reperfusion in HFD hearts. before ischemia. In this context, one possible mecha- Te same logic can be applied here: Te present fnding nism is the compromised GLUT4 translocation. Previous that an increased myocardial glucose uptake in response studies showed that GLUT4 is an important mediator to IRI was blocked by phlorizin may be the main mech- of enhanced glycolysis and maintaining ATP concentra- anism underlying the poor functional recovery and tion under various pathological conditions, including IRI increased myocardial injury after ischemia–reperfusion. [2, 8, 9, 12]. However, signifcant decreases in the sar- Tis aligns with our previous fndings showing that car- colemmal GLUT4 expression were observed in insulin- diac SGLT1 is involved in an important protective mech- resistant obese mouse hearts [13, 14], consistent with anism against IRI by replenishing ATP stores in ischemic the present fndings. Furthermore, we obtained the new cardiac tissues via enhanced glucose utilization [7]. Con- fnding that the GLUT4 upregulation in response to IRI sistent with a series of our studies, Connelly et al. showed insult was signifcantly attenuated in HFD mice, in asso- that oral administration of a dual SGLT1/2-inhibitor led ciation with the decreased myocardial glucose uptake to an impaired cardiac function after myocardial infarc- after IRI. Te impaired GLUT4 expression and transloca- tion in a rat model with left anterior descending coronary tion in diabetic subjects may involve not only disturbance artery ligation, while a selective SGLT2-inhibitor had no of insulin signaling but also increased membrane choles- signifcant efect on this condition [38]. Tese data indi- terol, reductions in membrane fuidity, and disruption cate that cardiac SGLT1 exerts protective efects against of caveolae and caveolin-3 [34]. Tese fndings may, at myocardial ischemia even in vivo, although they used a least in part, account for the impaired cardiac functional non-diabetic model rather than HFD-induced diabetic Yoshii et al. Cardiovasc Diabetol (2019) 18:85 Page 10 of 14

subjects. Furthermore, Kanwal et al. showed that SGLT1 fndings of the present study. Future investigations should inhibition by phlorizin abrogated the benefcial efect clarify the mechanisms underlying the regulation of the of ischemic preconditioning against IRI [39]. However, SGLT1 expression/activation in the cardiomyocytes in while cardiac SGLT1 does indeed exert favorable efects greater detail and identify the specifc regulators and/or during the acute phase of IRI, its chronic excessive acti- agents (namely, SGLT1-specifc agonists) that transiently vation has been reported to have unfavorable efects induce the expression/activation of cardiac SGLT1 dur- [20–23, 25], thus suggesting that the time course of SGLT ing IRI, which may lead to potential therapeutic applica- activation is critical for eliciting cardioprotective efects tions for the acute phase of IRI. [7]. Indeed, Li et al. very recently reported that the RNAi- A recent series of large-scale clinical trials has shown mediated knockdown of SGLT1 in cardiomyocytes is that SGLT2-inhibitors have profound benefts on reduc- protective against IRI by reducing oxidative stress [24], ing the risk of cardiovascular events in patients with type although they found a similar exacerbation of IRI when 2 diabetes [31–33]. In this context, assessing the direct using phlorizin. Tere may also be some compensa- efects of SGLT2-inhibitors on the heart in the current tory mechanisms activated in those mouse hearts by the experimental model has important clinical implications. permanent knock-down of cardiac SGLT1 from birth. We did not detect any SGLT2 expression, even in HFD Considering that the SGLT1 expression is chronically mouse hearts, which explains that tofoglifozin, ipragli- increased under diabetic conditions [19, 21], resulting in fozin, and canaglifozin, all of which are selective SGLT2- detrimental efects [21, 25, 40], it is important to attain inhibitors, did not afect the cardiac functional recovery the efects of SGLT1 on the cardiac function as well as or myocardial injury after IRI. To confrm the safety and energy metabolism during acute phase of IRI, particu- efcacy of these recently developed anti-diabetic agents, larly in insulin-resistant diabetic models. We found in the dosage used in the current study was almost fve- the present study that, under insulin-resistant conditions, fold higher than the blood concentrations after a single SGLT1 isoform may play a compensatory protective role oral dose of each SGLT2-inhibitor in clinical use. Tere during the loss of GLUT4-mediated glucose uptake, is growing evidence showing the potential cardioprotec- rather than provides detrimental impacts on the diabetic tive mechanisms of SGLT2-inhibitors, largely mediated myocardium [21, 25], at least during the acute phase of through (as systemic efects) decreasing the insulin resist- IRI. ance by reducing the body fat mass, modulating natriu- Despite the absence of changes in the SGLT1 expres- retic peptides, reducing arterial stifness, modulating sion, even under insulin-resistant conditions, our results sympathetic tone, exerting renoprotective efects, and (as still indicate the possible role of SGLT1 activation in the direct cardiac efects) reducing infammation, oxidative cardiomyocytes (in addition to its translocation/internal- stress, apoptosis, fbrosis, inhibiting sodium–hydrogen ization-mediated expression), considering that phlori- exchanger (NHE) [42–44]. Te neutral efect of SGLT2- zin exerted substantial efects on the cardiac functional inhibitors in the present study can be partly explained by recovery and injury as well as the myocardial glucose the fact that the current experimental model is an ex vivo uptake after ischemia–reperfusion in the HFD-induced isolated heart perfusion model, which eliminates the sys- obese mice. However, the precise regulatory mecha- temic efects of SGLT2-inhibitors. Meanwhile, a series nisms by which SGLT1 develops tolerance to IRI under of recent studies by Zuurbier et al. clearly showed that insulin-resistant conditions remain unclear. Te cardiac selective SGLT2-inhibitors directly inhibit cardiac NHE SGLT1 is thought to be activated by various factors, such [45–47], although empaglifozin did not signifcantly as insulin, AMP-activated protein kinase (AMPK), per- afect the cardiac functional recovery or myocardial sistent hyperglycemia, as well as ischemia–reperfusion injury after ischemia–reperfusion, which is consistent injury per se [18–21, 24, 40], all of which can also activate with our fndings; however, the authors did not examine GLUT4. And thus, a pathway specifc for cardiac SGLT1 any models of diet-induced obesity [45]. In addition, Lee activation has not yet been identifed. Hypoxia-induc- et al. reported that dapaglifozin attenuated cardiac fbro- ible factor 1α (HIF-1α), which is increased under dia- sis by regulating macrophage polarization [48], although betic as well as hypoxic conditions, is a major regulator most macrophage and other blood cells were washed out of the GLUT1 expression (although the GLUT1 expres- in our present study model of Langendorf. Intriguingly, sion was not signifcantly afected in the present study; they indicated that dapaglifozin induced compensatory Additional fle 1: Figs. S3 and S4). As such, it may play activation of SGLT1 as a cardio-protective mechanism, a potential role in regulating the SGLT1 activity in the leading to a reduction in oxidative stress. Based on these heart, although it has recently been shown to diminish previous fndings, it is possible that, in our current study, the SGLT1 and SGLT2 expression in kidneys [41], which selective SGLT2-inhibitors exerted direct cardioprotec- is the opposite of what we might expect based on the tive efects, although further studies will be needed in Yoshii et al. Cardiovasc Diabetol (2019) 18:85 Page 11 of 14

order to investigate the efects of those pharmacological GLUT1 were unchanged by phlorizin administration agents on cardiac NHE under diabetic insulin-resistant at all points of IRI under both diet conditions (Addi- conditions as well as on cardiac fbrosis through his- tional fle 1: Fig. S4). Likewise, previous studies by topathological analyses. Regardless, the present study other groups [22, 24] showed that there was no signif- investigated the transient local direct efects of SGLT2- cant diferences in the levels of membrane-associated inhibitors on the cardiac tissue during the acute phase of GLUT1 or GLUT4 in mice with cardiomyocyte-specifc IRI; our fndings therefore do not confict with the con- knockdown of SGLT1, suggesting that the expression/ sistent cardioprotective results from other in vivo experi- activity of SGLT1 and GLUTs may be regulated inde- mental studies [44, 49–56] or clinical trials [31–33], all of pendently, without any mutual interaction in the heart. which investigated the systemic efects of SGLT2-inhibi- It is also important to assess the expression of those tors over a longer period. Furthermore, the present fnd- glucose transporters in isolated cardiomyocytes, consid- ings suggest that potential mechanisms underlying the ering that not only myocardial SGLT1 and other glucose cardiovascular benefts of SGLT2-inhibitors may depend transporters but also those residing in cardiac fbroblasts largely on the systemic inter-organ network, including [40] and endothelial cells [67] may play a substantial role cardiorenal syndrome [57–59]. under diverse pathological conditions, such as ischemic GLUT1 is another major glucose transporter in the heart diseases as well as diabetes (insulin resistant condi- heart that is largely responsible for the basal cardiac tions), and the regulation of the expression of those glu- glucose transport in an insulin-independent manner [6, cose transporters might difer among cell types. 10]. Previous studies have shown that GLUT1 also exerts In this context, it would be interesting to clarify the important cardioprotective actions, mainly in chronic efects of SGLT1 inhibition during IRI in vivo in order pathological settings [15, 60]. Since GLUT1 is a facili- to examine its infuence on the systemic inter-organ tated energy-independent transporter that usually local- network as well as macrophages and other blood cells izes to the plasma membrane, its expression at the plasma circulating coronary arteries. On the other hand, the membrane is considered to directly refect its activity. Langendorf system is useful for evaluating the direct Given that we did not detect any signifcant change in the local efects of glucose transporters in the heart more GLUT1 expression at the plasma membrane, regardless clearly, as this system is not associated with changes in of the diet condition, GLUT1 may have a limited impact, systemic substrate metabolism or neurohumoral factors at least in the present experimental model. activated under pathological conditions, such as diabetes. Some studies suggested that 2-DG might be a rela- Given these limitations associated with the present tively poor substrate for SGLT1, although this glucose study, future investigations should conduct a series of analog has been used for the glucose uptake assay of experiments using a conditional (transient) cardiomyo- SGLTs in previous studies [7, 19, 61]. Tus, we can- cyte-specifc SGLT1 knockout model (rather than per- not completely deny the possibility that impairment manent knockdown from birth) in order to clarify the of cardiac glucose uptake by phlorizin may be due, at complete in vivo characterization of the relatively acute least in part, to the non-specifc inhibition of other glu- actions of SGLT1 in the cardiomyocytes. cose transporters. However, a series of previous stud- ies using phlorizin indicated that phlorizin does inhibit cardiac SGLT1 [25, 62–64]. Lambert et al. showed that Conclusions 200 µM of phlorizin reduced the myocardial intracellu- We determined that the impairment of the cardiac func- lar sodium concentration, which rules out any involve- tional recovery and increased myocardial injury associ- ment in GLUTs (namely, sodium-independent glucose ated with the reduction in the myocardial glucose uptake transporters) [21]. Furthermore, a relatively low dose by phlorizin during IRI were more pronounced in diet- of phlorizin was used in the present study compared induced obese mice hearts than in the hearts of mice fed to that used in the previous studies (200–10 mM) a normal diet, while no harmful efects induced by any [21, 25, 39, 62, 63], thus minimizing the non-specifc SGLT2-inhibitors were observed in the same setting. inhibitory efects of phlorizin on other glucose trans- Given that SGLT1 and GLUTs function independently porters. Finally, it is highly possible that phlorizin- without any mutual interaction, at least under the condi- perfusion specifcally inhibits cardiac SGLT1 in the tions in the present study, there is an increasing reliance current ex vivo Langendorf model, since non-specifc on SGLT1 for cardiac functional recovery after ischemia– inhibitory efects on glucose transporters are observed reperfusion via enhanced myocardial glucose utilization, only when phlorizin is broken down in the small intes- particularly under insulin-resistant conditions, in which tine to its aglycone form, phloretin [65, 66]. In fact, the GLUT4 upregulation in response to IRI is compromised. plasma membrane expression patterns of GLUT4 and Yoshii et al. Cardiovasc Diabetol (2019) 18:85 Page 12 of 14

Additional fle Funding This study was supported in part by grants-in-aid for the Ministry of Education, Culture, Sports, Science and Technology [JP17K09531] (to T. Nagoshi). Additional fle 1: Table S1. Baseline cardiac function of perfused hearts with or without SGLT2-inhibitors perfusion in HFD-fed mice. Table S2. Availability of data and materials Baseline cardiac function of perfused hearts with or without tofoglifo- The datasets used and/or analysed during the current study are available from zin perfusion in normal chow-fed mice. Fig. S1. The profles of other the corresponding author on reasonable request. hemodynamic parameters. The profles of RPP (A), LVEDP (B), and positive and negative dp/dt (C) measured at the indicated time points during Ethics approval and consent to participate ischemia–reperfusion in NFD (open black square; n 6), NFD phlorizin- All animal procedures conformed to the National Institutes of Health Guide = perfused (flled black square; n 6), HFD (open pink square; n 8), and for the Care and Use of Laboratory Animals and were approved by the Animal = = HFD phlorizin-perfused (flled pink square; n 8) hearts are shown. Research Committee at the Jikei University School of Medicine (2016-038C4). *P < 0.05 and **P < 0.01 versus NFD group at =each time point; †P < 0.05 and ‡P < 0.01 versus HFD group at each time point. min, minutes; RPP, rate Consent for publication pressure product; LVEDP, left ventricular end-diastolic pressure. Fig. S2. Not applicable. SGLT2 mRNA was not detected in the heart from both NFD and HFD mice. The quantitative reverse transcription polymerase chain reaction (QRT- Competing interests PCR) data indicating the SGLT2 gene expression levels in the hearts from Michihiro Yoshimura reports that there are any relevant conficts of interest. either NFD (A) or HFD (B), or in the mouse intestine as the negative control Outside the submitted work, there are personal fees from Mochida Pharma- and in the kidney as the positive control (C) (n 3 each). (D) The QRT-PCR ceutical Co., Ltd, personal fees from MSD K.K., grants and personal fees from = data were normalized to GAPDH. The data are shown as the fold change Kyowa Hakko Kirin Co., Ltd., grants and personal fees from Pfzer Japan Inc., normalized to the levels found in the kidney (C). Fig. S3. Expression of grants and personal fees from Daiichi Sankyo Healthcare Co., Ltd., grants and GLUT1 in the murine hearts during ischemia–reperfusion. Representative personal fees from Sumitomo Dainippon Pharma Co., Ltd., grants from Nippon immunoblots of GLUT1 in the plasma membrane fraction from the murine Boehringer Ingelheim Co., Ltd., grants from Otsuka Pharmaceutical Co., Ltd., perfused hearts at baseline period measured at the end of the 10-minute grants from Teijin Pharma Limited, personal fees from Fujiyakuhin Co., Ltd., pre-ischemia perfusion (A), and before and after IRI (B) are shown. (C) grants and personal fees from AstraZeneca K.K., personal fees from Taisho Toy- Densitometric quantitation normalized to the level of GLUT1 expression in ama Pharmaceutical Co., Ltd., grants from Astellas Pharma Inc., personal fees NFD hearts before IRI is shown (NFD or HFD without IRI; n 5 each, with from Mitsubishi Tanabe Pharma Corporation, grants from Shionogi & Co., Ltd., = IRI; n 3 each). In both (A) and (B), immunoblots of ­Na+/K+ ATPase from grants from Ono Pharmaceutical Co., Ltd. Tomohisa Nagoshi also reports that = the same membrane are shown as a loading control for the membrane there are any relevant conficts of interest. Outside the submitted work, there fraction. Fig. S4. Expression of GLUT4, SGLT1 and GLUT1 in murine are grants from Daiichi Sankyo Healthcare Co., Ltd., from Nippon Boehringer hearts during ischemia–reperfusion with or without phlorizin-perfusion. Ingelheim Co., Ltd., and from Mitsubishi Tanabe Pharma Corporation. Representative immunoblots of GLUT4, SGLT1 and GLUT1 in the plasma membrane fraction from the murine perfused hearts before and after Received: 22 March 2019 Accepted: 23 June 2019 IRI with or without phlorizin-perfuion (A) are shown. The immunoblot of ­Na+/K+ ATPase from the same membrane are shown as a loading control for the membrane fraction. (B) Densitometric quantitation normalized to the level of either GLUT4, SGLT1 or GLUT1 expression in NFD hearts before IRI are shown (n 3 each). *P < 0.05, **P < 0.01 versus NFD hearts without = # phlorizin perfusion before IRI; P < 0.05 versus NFD hearts with phlorizin References perfusion before IRI; ‡P < 0.01 versus NFD hearts without phlorizin perfu- § 1. Depre C, Vanoverschelde JL, Taegtmeyer H. Glucose for the heart. Circula- sion after IRI; P < 0.05 versus NFD hearts with phlorizin perfusion after IRI. tion. 1999;99(4):578–88. 2. Tian R, Abel ED. Responses of GLUT4-defcient hearts to ischemia under- score the importance of glycolysis. Circulation. 2001;103(24):2961–6. Abbreviations 3. Stanley WC, Recchia FA, Lopaschuk GD. Myocardial substrate metabolism 2-DG: 2-deoxy-d-glucose; AAR:​ area at risk; AMPK: AMP-activated protein in the normal and failing heart. Physiol Rev. 2005;85(3):1093–129. kinase; ATP: adenosine triphosphate; AUC​: area under the curve; Cana: 4. Nagoshi T, Yoshimura M, Rosano GM, Lopaschuk GD, Mochizuki S. canaglifozin; CPK: creatine phosphokinase; DMSO: dimethylsulfoxide; dp/ Optimization of cardiac metabolism in heart failure. Curr Pharm Des. dt : maximum rate of contraction; dp/dt : maximum rate of relaxa+- max min 2011;17:3846–53. tion; GAPDH: glyceraldehyde 3-phosphate− dehydrogenase; GLUT: facilitated 5. Heywood SE, Richart AL, Henstridge DC, Alt K, Kiriazis H, Zammit C, Carey glucose transporter; HFD: high-fat diet; HIF-1α: hypoxia-inducible factor 1α; AL, Kammoun HL, Delbridge LM, Reddy M, et al. High-density lipoprotein HR: heart rate; I: ischemia; Ipra: ipraglifozin; IRI: ischemia–reperfusion injury; delivered after myocardial infarction increases cardiac glucose uptake LVDP: left ventricular developed pressure; LVEDP: left ventricular end-diastolic and function in mice. Sci Transl Med. 2017;9(411):e6084. pressure; LVSP: left ventricular systolic pressure; min: minutes; NFD: normal-fat 6. Szablewski L. Glucose transporters in healthy heart and in cardiac disease. diet; QRT-PCR: quantitative reverse transcription polymerase chain reac- Int J Cardiol. 2017;230:70–5. tion; R: reperfusion; RNA: ribonucleic acid; RPP: rate pressure product; SGLT: 7. Kashiwagi Y, Nagoshi T, Yoshino T, Tanaka TD, Ito K, Harada T, Takahashi H, sodium–glucose cotransporter; SGLT2i: SGLT2-inhibitor; Tofo: tofoglifozin; TTC​: Ikegami M, Anzawa R, Yoshimura M. Expression of SGLT1 in human hearts triphenyltetrazolium chloride. and impairment of cardiac glucose uptake by phlorizin during ischemia– reperfusion injury in mice. PLoS ONE. 2015;10(6):e0130605. Acknowledgements 8. Matsui T, Tao J, del Monte F, Lee KH, Li L, Picard M, Force TL, Franke TF, The authors are grateful to Sanof K.K. (Tokyo), Astellas Pharma Inc. (Ibaraki), Hajjar RJ, Rosenzweig A. Akt activation preserves cardiac function and and Mitsubishi Tanabe Pharma Co. (Osaka) for supplying tofoglifozin, ipragli- prevents injury after transient cardiac ischemia in vivo. Circulation. fozin, and canaglifozin, respectively. 2001;104(3):330–5. 9. Matsui T, Nagoshi T, Hong EG, Luptak I, Hartil K, Li L, Gorovits N, Charron Authors’ contributions MJ, Kim JK, Tian R, et al. Efects of chronic Akt activation on glucose AY, TN, and MY conceived and designed the study, contributed to the writing uptake in the heart. Am J Physiol Endocrinol Metab. 2006;290(5):E789–97. of the manuscript. AY, TN, YK, HK, YT, and YO conducted experiments, contrib- 10. Shao D, Tian R. Glucose transporters in cardiac metabolism and hypertro- uted to the acquisition and interpretation of data. KI, TY, and TDT analysed and phy. Compr Physiol. 2015;6(1):331–51. interpreted the data, and critically revised the manuscript. All authors read and 11. Nagoshi T, Matsui T, Aoyama T, Leri A, Anversa P, Li L, Ogawa W, Del Monte approved the fnal manuscript. F, Gwathmey JK, Grazette L, et al. PI3K rescues the detrimental efects of Yoshii et al. Cardiovasc Diabetol (2019) 18:85 Page 13 of 14

chronic Akt activation in the heart during ischemia/reperfusion injury. J obesity on infammation and remodeling after myocardial infarction. Am Clin Investig. 2005;115(8):2128–38. J Physiol Heart Circ Physiol. 2006;291(5):H2504–14. 12. Wende AR, Kim J, Holland WL, Wayment BE, O’Neill BT, Tuinei J, Brahma 30. Aoyagi T, Higa JK, Aoyagi H, Yorichika N, Shimada BK, Matsui T. Car- MK, Pepin ME, McCrory MA, Luptak I, et al. Glucose transporter diac mTOR rescues the detrimental efects of diet-induced obesity in 4-defcient hearts develop maladaptive hypertrophy in response to the heart after ischemia–reperfusion. Am J Physiol Heart Circ Physiol. physiological or pathological stresses. Am J Physiol Heart Circ Physiol. 2015;308(12):H1530–9. 2017;313(6):H1098–108. 31. Zinman B, Wanner C, Lachin JM, Fitchett D, Bluhmki E, Hantel S, 13. Wright JJ, Kim J, Buchanan J, Boudina S, Sena S, Bakirtzi K, Ilkun O, Mattheus M, Devins T, Johansen OE, Woerle HJ, et al. Empaglifozin, Theobald HA, Cooksey RC, Kandror KV, et al. Mechanisms for increased cardiovascular outcomes, and mortality in type 2 diabetes. N Engl J Med. myocardial fatty acid utilization following short-term high-fat feeding. 2015;373(22):2117–28. Cardiovasc Res. 2009;82(2):351–60. 32. Neal B, Perkovic V, Mahafey KW, de Zeeuw D, Fulcher G, Erondu N, Shaw 14. Cook SA, Varela-Carver A, Mongillo M, Kleinert C, Khan MT, Leccisotti L, W, Law G, Desai M, Matthews DR, et al. Canaglifozin and cardiovascular Strickland N, Matsui T, Das S, Rosenzweig A, et al. Abnormal myocardial and renal events in type 2 diabetes. N Engl J Med. 2017;377(7):644–57. insulin signalling in type 2 diabetes and left-ventricular dysfunction. Eur 33. Wiviott SD, Raz I, Bonaca MP, Mosenzon O, Kato ET, Cahn A, Silverman MG, Heart J. 2010;31(1):100–11. Zelniker TA, Kuder JF, Murphy SA, et al. Dapaglifozin and cardiovascular 15. Liao R, Jain M, Cui L, D’Agostino J, Aiello F, Luptak I, Ngoy S, Mortensen outcomes in type 2 diabetes. N Engl J Med. 2019;380(4):347–57. RM, Tian R. Cardiac-specifc overexpression of GLUT1 prevents the 34. Russell J, Du Toit EF, Peart JN, Patel HH, Headrick JP. Myocyte membrane development of heart failure attributable to pressure overload in mice. and microdomain modifcations in diabetes: determinants of ischemic Circulation. 2002;106(16):2125–31. tolerance and cardioprotection. Cardiovasc Diabetol. 2017;16(1):155. 16. Zhou L, Cryan EV, D’Andrea MR, Belkowski S, Conway BR, Demarest KT. 35. Jia G, Hill MA, Sowers JR. Diabetic cardiomyopathy. Circ Res. Human cardiomyocytes express high level of Na /glucose cotransporter 2018;122(4):624–38. 1 (SGLT1). J Cell Biochem. 2003;90(2):339–46. + 36. Chong CR, Clarke K, Levelt E. Metabolic remodeling in diabetic cardio- 17. Wright EM, Loo DD, Hirayama BA. Biology of human sodium glucose myopathy. Cardiovasc Res. 2017;113(4):422–30. transporters. Physiol Rev. 2011;91(2):733–94. 37. Bugger H, Abel ED. Rodent models of diabetic cardiomyopathy. Dis 18. Di Franco A, Cantini G, Tani A, Coppini R, Zecchi-Orlandini S, Raimondi L, Model Mech. 2009;2(9–10):454–66. Luconi M, Mannucci E. Sodium-dependent glucose transporters (SGLT) 38. Connelly KA, Zhang Y, Desjardins J-F, Thai K, Gilbert RE. Dual inhibition in human ischemic heart: a new potential pharmacological target. Int J of sodium–glucose linked cotransporters 1 and 2 exacerbates cardiac Cardiol. 2017;243:86–90. dysfunction following experimental myocardial infarction. Cardiovasc 19. Banerjee SK, McGafn KR, Pastor-Soler NM, Ahmad F. SGLT1 is a novel Diabetol. 2018;17(1):99. cardiac glucose transporter that is perturbed in disease states. Cardiovasc 39. Kanwal A, Nizami HL, Mallapudi S, Putcha UK, Mohan GK, Banerjee SK. Res. 2009;84(1):111–8. Inhibition of SGLT1 abrogates preconditioning-induced cardioprotection 20. Banerjee SK, Wang DW, Alzamora R, Huang XN, Pastor-Soler NM, Hallows against ischemia–reperfusion injury. Biochem Biophys Res Commun. KR, McGafn KR, Ahmad F. SGLT1, a novel cardiac glucose transporter, 2016;472(2):392–8. mediates increased glucose uptake in PRKAG2 cardiomyopathy. J Mol 40. Meng L, Uzui H, Guo H, Tada H. Role of SGLT1 in high glucose level- Cell Cardiol. 2010;49(4):683–92. induced MMP-2 expression in human cardiac fbroblasts. Mol Med Rep. 21. Lambert R, Srodulski S, Peng X, Margulies KB, Despa F, Despa S. 2018;17(5):6887–92. Intracellular Na concentration ([Na ]i) is elevated in diabetic 41. Zapata-Morales JR, Galicia-Cruz OG, Franco M, Martinez YMF. Hypoxia- hearts due to enhanced+ Na -glucose+ cotransport. J Am Heart Assoc. inducible factor-1alpha (HIF-1alpha) protein diminishes sodium glucose 2015;4(9):e002183. + transport 1 (SGLT1) and SGLT2 protein expression in renal epithelial 22. Ramratnam M, Sharma RK, D’Auria S, Lee SJ, Wang D, Huang XY, Ahmad tubular cells (LLC-PK1) under hypoxia. J Biol Chem. 2014;289(1):346–57. F. Transgenic knockdown of cardiac sodium/glucose cotransporter 42. Heerspink HJ, Perkins BA, Fitchett DH, Husain M, Cherney DZ. Sodium glu- 1 (SGLT1) attenuates PRKAG2 cardiomyopathy, whereas transgenic cose cotransporter 2 inhibitors in the treatment of diabetes: cardiovas- overexpression of cardiac SGLT1 causes pathologic hypertrophy and cular and kidney efects, potential mechanisms and clinical applications. dysfunction in mice. J Am Heart Assoc. 2014;3(4):e000899. Circulation. 2016;134(10):752–72. 23. Matsushita N, Ishida N, Ibi M, Saito M, Sanbe A, Shimojo H, Suzuki S, 43. Lahnwong S, Chattipakorn SC, Chattipakorn N. Potential mechanisms Koepsell H, Takeishi Y, Morino Y, et al. Chronic pressure overload induces responsible for cardioprotective efects of sodium–glucose co-trans- cardiac hypertrophy and fbrosis via increases in SGLT1 and IL-18 gene porter 2 inhibitors. Cardiovasc Diabetol. 2018;17(1):101. expression in mice. Int Heart J. 2018;59:1123–33. 44. Chin KL, Ofori-Asenso R, Hopper I, von Lueder TG, Reid CM, Zoungas 24. Li Z, Agrawal V, Ramratnam M, Sharma RK, D’Auria S, Sincoular A, Jakubiak S, Wang BH, Liew D. Potential mechanisms underlying the car- M, Music ML, Kutschke WJ, Huang XN, et al. Cardiac sodium–glucose co- diovascular benefts of sodium glucose cotransporter 2 inhibitors: transporter 1 (SGLT1) is a novel mediator of ischemia/reperfusion injury. a systematic review of data from preclinical studies. Cardiovasc Res. Cardiovasc Res. 2019;56:56. https​://doi.org/10.1093/cvr/cvz03​7 (in Press). 2019;115(2):266–76. 25. Balteau M, Tajeddine N, de Meester C, Ginion A, Des Rosiers C, Brady 45. Uthman L, Nederlof R, Eerbeek O, Baartscheer A, Schumacher C, Buch- NR, Sommereyns C, Horman S, Vanoverschelde JL, Gailly P, et al. NADPH holtz N, Hollmann MW, Coronel R, Weber NC, Zuurbier CJ. Delayed oxidase activation by hyperglycaemia in cardiomyocytes is inde- ischemic contracture onset by empaglifozin associates with NHE-1 inhi- pendent of glucose metabolism but requires SGLT1. Cardiovasc Res. bition and is dependent on insulin in isolated mouse hearts. Cardiovasc 2011;92(2):237–46. Res. 2019. https​://doi.org/10.1093/cvr/cvz00​4 (in Press). 26. Yoshino T, Nagoshi T, Anzawa R, Kashiwagi Y, Ito K, Katoh D, Fujisaki M, 46. Uthman L, Baartscheer A, Bleijlevens B, Schumacher CA, Fiolet JWT, Koe- Kayama Y, Date T, Hongo K, et al. Preconditioning actions of aldosterone man A, Jancev M, Hollmann MW, Weber NC, Coronel R, et al. Class efects through p38 signaling modulation in isolated rat hearts. J Endocrinol. of SGLT2 inhibitors in mouse cardiomyocytes and hearts: inhibition of 2014;222(2):289–99. Na( )/H( ) exchanger, lowering of cytosolic Na( ) and vasodilation. 27. Nagoshi T, Date T, Fujisaki M, Yoshino T, Sekiyama H, Ogawa K, Kayama Y, Diabetologia.+ + 2018;61(3):722–6. + Minai K, Komukai K, Ogawa T, et al. Biphasic action of aldosterone on Akt 47. Baartscheer A, Schumacher CA, Wust RC, Fiolet JW, Stienen GJ, Coronel signaling in cardiomyocytes. Horm Metab Res. 2012;44:931–7. R, Zuurbier CJ. Empaglifozin decreases myocardial cytoplasmic Na 28. Kimura H, Nagoshi T, Yoshii A, Kashiwagi Y, Tanaka Y, Ito K, Yoshino T, through inhibition of the cardiac Na /H exchanger in rats and rabbits.+ Tanaka TD, Yoshimura M. The thermogenic actions of natriuretic peptide Diabetologia. 2016;60(3):568–73. + + in brown adipocytes: the direct measurement of the intracellular tem- 48. Lee TM, Chang NC, Lin SZ. Dapaglifozin, a selective SGLT2 inhibitor, perature using a fuorescent thermoprobe. Sci Rep. 2017;7(1):12978. attenuated cardiac fbrosis by regulating the macrophage polariza- 29. Thakker GD, Frangogiannis NG, Bujak M, Zymek P, Gaubatz JW, Reddy AK, tion via STAT3 signaling in infarcted rat hearts. Free Radic Biol Med. Tafet G, Michael LH, Entman ML, Ballantyne CM. Efects of diet-induced 2017;104:298–310. Yoshii et al. Cardiovasc Diabetol (2019) 18:85 Page 14 of 14

49. Lin B, Koibuchi N, Hasegawa Y, Sueta D, Toyama K, Uekawa K, Ma M, Naka- 59. Sano M, Goto S. Possible mechanism of hematocrit elevation by sodium gawa T, Kusaka H, Kim-Mitsuyama S. Glycemic control with empaglifozin, glucose cotransporter 2 inhibitors and associated benefcial renal and a novel selective SGLT2 inhibitor, ameliorates cardiovascular injury and cardiovascular efects. Circulation. 2019;139(17):1985–7. cognitive dysfunction in obese and type 2 diabetic mice. Cardiovasc 60. Yan J, Young ME, Cui L, Lopaschuk GD, Liao R, Tian R. Increased glucose Diabetol. 2014;13(1):148. uptake and oxidation in mouse hearts prevent high fatty acid oxida- 50. Kusaka H, Koibuchi N, Hasegawa Y, Ogawa H, Kim-Mitsuyama S. Empagli- tion but cause cardiac dysfunction in diet-induced obesity. Circulation. fozin lessened cardiac injury and reduced visceral adipocyte hypertro- 2009;119(21):2818–28. phy in prediabetic rats with metabolic syndrome. Cardiovasc Diabetol. 61. Kanwal A, Singh SP, Grover P, Banerjee SK. Development of a cell-based 2016;15(1):157. nonradioactive glucose uptake assay system for SGLT1 and SGLT2. Anal 51. Joubert M, Jagu B, Montaigne D, Marechal X, Tesse A, Ayer A, Dollet L, Le Biochem. 2012;429(1):70–5. May C, Toumaniantz G, Manrique A, et al. The sodium–glucose cotrans- 62. von Lewinski D, Gasser R, Rainer PP, Huber MS, Wilhelm B, Roessl U, porter 2 inhibitor dapaglifozin prevents cardiomyopathy in a diabetic Haas T, Wasler A, Grimm M, Bisping E, et al. Functional efects of glu- lipodystrophic mouse model. Diabetes. 2017;66(4):1030–40. cose transporters in human ventricular myocardium. Eur J Heart Fail. 52. Habibi J, Aroor AR, Sowers JR, Jia G, Hayden MR, Garro M, Barron B, May- 2010;12(2):106–13. oux E, Rector RS, Whaley-Connell A, et al. Sodium glucose transporter 2 63. von Lewinski D, Rainer PP, Gasser R, Huber MS, Khafaga M, Wilhelm B, (SGLT2) inhibition with empaglifozin improves cardiac diastolic function Haas T, Machler H, Rossl U, Pieske B. Glucose-transporter-mediated posi- in a female rodent model of diabetes. Cardiovasc Diabetol. 2017;16(1):9. tive inotropic efects in human myocardium of diabetic and nondiabetic 53. Verma S, Rawat S, Ho KL, Wagg CS, Zhang L, Teoh H, Dyck JE, Uddin GM, patients. Metabolism. 2010;59(7):1020–8. Oudit GY, Mayoux E, et al. Empaglifozin increases cardiac energy produc- 64. Ohgaki R, Wei L, Yamada K, Hara T, Kuriyama C, Okuda S, Ueta K, Shiotani tion in diabetes. JACC Basic Transl Sci. 2018;3(5):575–87. M, Nagamori S, Kanai Y. Interaction of the sodium/glucose cotransporter 54. Tanajak P, Sa-Nguanmoo P, Sivasinprasasn S, Thummasorn S, Siri-Angkul (SGLT) 2 inhibitor canaglifozin with SGLT1 and SGLT2. J Pharmacol Exp N, Chattipakorn SC, Chattipakorn N. Cardioprotection of dapaglifozin Ther. 2016;358(1):94–102. and vildagliptin in rats with cardiac ischemia–reperfusion injury. J Endo- 65. Han S, Hagan DL, Taylor JR, Xin L, Meng W, Biller SA, Wetterau JR, crinol. 2018;236(2):69–84. Washburn WN, Whaley JM. Dapaglifozin, a selective SGLT2 inhibitor, 55. Adingupu DD, Gopel SO, Gronros J, Behrendt M, Sotak M, Miliotis T, Dahl- improves glucose homeostasis in normal and diabetic rats. Diabetes. qvist U, Gan LM, Jonsson-Rylander AC. SGLT2 inhibition with empaglifo- 2008;57(6):1723–9. zin improves coronary microvascular function and cardiac contractility in 66. Hawley SA, Ford RJ, Smith BK, Gowans GJ, Mancini SJ, Pitt RD, Day EA, Salt prediabetic ob/ob( / ) mice. Cardiovasc Diabetol. 2019;18(1):16. IP, Steinberg GR, Hardie DG. The Na /glucose cotransporter inhibitor 56. Li C, Zhang J, Xue M,− Li− X, Han F, Liu X, Xu L, Lu Y, Cheng Y, Li T, et al. SGLT2 canaglifozin activates AMPK by inhibiting+ mitochondrial function and inhibition with empaglifozin attenuates myocardial oxidative stress and increasing cellular AMP levels. Diabetes. 2016;65(9):2784–94. fbrosis in diabetic mice heart. Cardiovasc Diabetol. 2019;18(1):15. 67. Elfeber K, Stumpel F, Gorboulev V, Mattig S, Deussen A, Kaissling B, 57. Perkovic V, Jardine MJ, Neal B, Bompoint S, Heerspink HJL, Charytan Koepsell H. Na( )-d-glucose cotransporter in muscle capillaries increases DM, Edwards R, Agarwal R, Bakris G, Bull S, et al. Canaglifozin and glucose permeability.+ Biochem Biophys Res Commun. 2004;314(2):301–5. renal outcomes in type 2 diabetes and nephropathy. N Engl J Med. 2019;380(24):2295–306. 58. Wanner C, Inzucchi SE, Lachin JM, Fitchett D, von Eynatten M, Mattheus Publisher’s Note M, Johansen OE, Woerle HJ, Broedl UC, Zinman B, et al. Empaglifozin Springer Nature remains neutral with regard to jurisdictional claims in pub- and progression of kidney disease in type 2 diabetes. N Engl J Med. lished maps and institutional afliations. 2016;375(4):323–34.

Ready to submit your research ? Choose BMC and benefit from:

• fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year

At BMC, research is always in progress.

Learn more biomedcentral.com/submissions